Since the network device hotplug is an asynchronous operation,
it's possible that the hotplug operation had returned, but
the network device hasn't ready in guest, thus it's better to
retry on this operation to wait until the device ready in guest.
Signed-off-by: Fupan Li <fupan.lfp@antgroup.com>
We have here either /dev/vfio/<num> or /dev/vfio/devices/vfio<num>,
for IOMMUFD format /dev/vfio/devices/vfio<num>, strip "vfio" prefix
/dev/vfio/123 - basename "123" - vfioNum = "123" - cdi.k8s.io/vfio123
/dev/vfio/devices/vfio123 - basename "vfio123" - strip - vfioNum = "123" - cdi.k8s.io/vfio123
Signed-off-by: Zvonko Kaiser <zkaiser@nvidia.com>
This change crystallizes and simplifies the current handling of /dev
hostPath mounts with virtually no functional change.
Before this change:
- If a mount DESTINATION is in /dev and it is a non-regular file on the HOST,
the shim passes the OCI bind mount as is to the guest (e.g.
/dev/kmsg:/dev/kmsg). The container rightfully sees the GUEST device.
- If the mount DESTINATION does not exist on the host, the shim relies on
k8s/containerd to automatically create a directory (ie. non-regular file) on
the HOST. The shim then also passes the OCI bind mount as is to the guest. The
container rightfully sees the GUEST device.
- For other /dev mounts, the shim passes the device major/minor to the guest
over virtio-fs. The container rightfully sees the GUEST device.
After this change:
- If a mount SOURCE is in /dev and it is a non-regular file on the HOST,
the shim passes the OCI bind mount as is to the guest. The container
rightfully sees the GUEST device.
- The shim does not anymore rely on k8s/containerd to create missing mount
directories. Instead it explicitely handles missing mount SOURCES, and
treats them like the previous bullet point.
- The shim no longer uses virtio-fs to pass /dev device major/minor to the
guest, instead it passes the OCI bind mount as is.
Signed-off-by: Aurélien Bombo <abombo@microsoft.com>
On s390x, QEMU fails if maxmem is set to 0:
```
invalid value of maxmem: maximum memory size (0x0) must be at least the initial memory size
```
This commit sets maxmem to the initial memory size for s390x when hotplug is disabled,
resolving the error while still ensuring that memory hotplug remains off.
Signed-off-by: Hyounggyu Choi <Hyounggyu.Choi@ibm.com>
The setting '-m xM,slots=y,maxmem=zM' where maxmem is from
the host's memory capacity is failing with confidential VMs
on hosts having 1T+ of RAM.
slots/maxmem are necessary for setups where the container
memory is hotplugged to the VM during container creation based
on createContainer info.
This is not the case with CoCo since StaticResourceManagement
is enabled and memory hotplug flows have not been checked.
To avoid unexpeted errors with maxmem, disable slots/maxmem
in case ConfidentialGuest is requested.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
This commit introduces changes to parse the PciDeviceInfo received
in response payload when adding a network device to the VM with cloud
hypervisor. When hotplugging a network device for a given endpoint,
it rightly sets the PciPath of the plugged-in device in the endpoint.
In calls like virtcontainers/sandbox.go:AddInterface, the later call
to agent sends the pci info for uevents (instead of empty value) to
rightly update the interfaces instead of failing with `Link not found`
Signed-off-by: Sumedh Alok Sharma <sumsharma@microsoft.com>
Since the cloud hypervisor's resize vCPU is an asynchronous operation,
it's possible that the previous resize operation hasn't completed when
the request is sent, causing the current call to return an error.
Therefore, several retries can be performed to avoid this error.
Signed-off-by: Fupan Li <fupan.lfp@antgroup.com>
```
Experimental fw_cfg Device Support
This feature enables passing configuration data and files, such as VM
boot configurations (kernel, kernel cmdline, e820 memory map, and ACPI
tables), from the host to the guest. (#7117)
Experimental ivshmem Device Support
Support for inter-VM shared memory has been added. For more information,
please refer to the ivshmem documentation. (#6703)
Firmware Boot Support on riscv64
In addition to direct kernel boot, firmware boot support has been added
on riscv64 hosts. (#7249)
Increased vCPU Limit on x86_64/kvm
The maximum number of supported vCPUs on x86_64 hosts using KVM has been
raised from 254 to 8192. (#7299)
Improved Block Performance with Small Block Sizes
Performance for virtio-blk with small block sizes (16KB and below)
is enhanced via submitting async IO requests in batches. (#7146)
Faster VM Pause Operation
The VM pause operation now is significantly faster particularly for VMs
with a large number of vCPUs. (#7290)
Updated Documentation on Windows Guest Support
Our Windows documentation now includes instructions to run Windows 11
guests, in addition to Windows Server guests. (#7218)
Policy on AI Generated Code
We will decline any contributions known to contain contents generated or
derived from using Large Language Models (LLMs). Details can be found
in our contributing documentation. (#7162)
Removed SGX Support
The SGX support has been removed, as announced in the deprecation notice two
release cycles ago. (#7093)
Notable Bug Fixes
Seccomp filter fixes with glibc v2.42 (#7327)
Various fixes related to (#7331, #7334, #7335)
```
From https://github.com/cloud-hypervisor/cloud-hypervisor/releases/tag/v48.0
Signed-off-by: Fabiano Fidêncio <fabiano@fidencio.org>
ef642fe890 added a special case to avoid
moving cgroups that are on the "default" slice in case of deletion.
However, this special check should be done in the Parent() method
instead, which ensures that the default resource controller ID is
returned, instead of ".".
Fixes: #11599
Signed-off-by: Fabiano Fidêncio <fidencio@northflank.com>
Similar to what we've done for Cloud Hypervisor in the commit
9f76467cb7, we're backporting a runtime-rs
feature that would be benificial to have as part of the go runtime.
This allows users to use virito-balloon for the hypervisor to reclaim
memory freed by the guest.
Signed-off-by: Fabiano Fidêncio <fidencio@northflank.com>
As part of the go 1.24.6 bump there are errors about the incorrect
use of a errorf, so switch to the non-formatting version, or add
the format string as appropriate
Signed-off-by: stevenhorsman <steven@uk.ibm.com>
Let's rename the runtime-rs initdata annotation from
`io.katacontainers.config.runtime.cc_init_data` to
`io.katacontainers.config.hypervisor.cc_init_data`.
Rationale:
- initdata itself is a hypervisor-specific feature
- the new name aligns with the annotation handling logic:
c92bb1aa88/src/libs/kata-types/src/annotations/mod.rs (L514-L968)
This commit updates the annotation for go-runtime and tests accordingly.
Signed-off-by: Hyounggyu Choi <Hyounggyu.Choi@ibm.com>
Dependening on the platform configuration, users might want to
set a more secure policy than the QEMU default.
Signed-off-by: Paul Meyer <katexochen0@gmail.com>
When hot-plugging CPUs on QEMU, we send a QMP command with JSON
arguments. QEMU 9.2 recently became more strict[1] enforcing the
JSON schema for QMP parameters. As a result, running Kata Containers
with QEMU 9.2 results in a message complaining that the core-id
parameter is expected to be an integer:
```
qmp hotplug cpu, cpuID=cpu-0 socketID=1, error:
QMP command failed:
Invalid parameter type for 'core-id', expected: integer
```
Fix that by changing the core-id, socket-id and thread-id to be
integer values.
[1]: be93fd5372Fixes: #11633
Signed-off-by: Christophe de Dinechin <dinechin@redhat.com>
The generated Go bindings for the agent are out of date. This commit
was produced by running
src/agent/src/libs/protocols/hack/update-generated-proto.sh with
protobuf compiler versions matching those of the last run, according to
the generated code comments.
Since there are new RPC methods, those needed to be added to the
HybridVSockTTRPCMockImp.
Signed-off-by: Markus Rudy <mr@edgeless.systems>
Re-generates the client code against Cloud Hypervisor v47.0.
Note: The client code of cloud-hypervisor's OpenAPI is automatically
generated by openapi-generator.
Signed-off-by: Ruoqing He <heruoqing@iscas.ac.cn>
Previously, the rootlessDir variable in `src/runtime/virtcontainers/pkg/rootless.go` was initialized at
package load time using `os.Getenv("XDG_RUNTIME_DIR")`. However, in rootless
VMM mode, the correct value of $XDG_RUNTIME_DIR is set later during runtime
using os.Setenv(), so rootlessDir remained empty.
This patch defers the initialization of rootlessDir until the first call
to `GetRootlessDir()`, ensuring it always reflects the current environment
value of $XDG_RUNTIME_DIR.
Fixes: #11526
Signed-off-by: stevenfryto <sunzitai_1832@bupt.edu.cn>
Removing runtime SEV functionality,
such as the kbs, ovmf, VMSA handling,
and SEV configs as part of deprecating
SEV from kata.
Co-authored-by: Adithya Krishnan Kannan <AdithyaKrishnan.Kannan@amd.com>
Signed-off-by: Arvind Kumar <arvinkum@amd.com>
To better support containerd 2.1 and later versions, remove the
hardcoded `layer.erofs` and instead parse `/proc/mounts` to obtain the
real mount source (and `/sys/block/loopX/loop/backing_file` if needed).
If the mount source doesn't end with `layer.erofs`, it should be marked
as unsupported, as it may be a filesystem meta file generated by later
containerd versions for the EROFS flattened filesystem feature.
Also check whether the filesystem type is `overlay` or not, since the
containerd mount manager [1] may change it after being introduced.
[1] https://github.com/containerd/containerd/issues/11303
Fixes: f63ec50ba3 ("runtime: Add EROFS snapshotter with block device support")
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Fixes a confusing log message shown when Virtio-FS is disabled.
Previously we logged “The virtiofsd had stopped” regardless of whether Virtio-FS was actually enabled or not.
Signed-off-by: Paweł Bęza <pawel.beza99@gmail.com>
Allow users to build using DEFDISABLEIMAGENVDIMM=true if they want to
set disable_image_nvdimm=true in configuration-clh.toml.
disable_image_nvdimm=false is the default config value.
Also, use virtio-blk instead of nvdimm if disable_image_nvdimm=true in
configuration-clh.toml.
Signed-off-by: Dan Mihai <dmihai@microsoft.com>
Currently, when a new sandbox resource controller is created with cgroupsv2 and sandbox_cgroup_only is disabled,
the cgroup management falls back to cgroupfs. During deletion, `IsSystemdCgroup` checks if the path contains `:`
and tries to delete the cgroup via systemd. However, the cgroup was originally set up via cgroupfs and this process
fails with `lstat /sys/fs/cgroup/kubepods.slice/kubepods-besteffort.slice/....scope: no such file or directory`.
This patch updates the deletion logic to take in to account the sandbox_cgroup_only=false option and in this case uses
the cgroupfs delete.
Fixes: #11036
Signed-off-by: Champ-Goblem <cameron@northflank.com>
This enables guest pull via config, without the need of any external
snapshotter. When the config enables runtime.experimental_force_guest_pull, instead of
relying on annotations to select the way to share the root FS, we always
use guest pull.
Co-authored-by: Markus Rudy <mr@edgeless.systems>
Signed-off-by: Paul Meyer <katexochen0@gmail.com>
Fixes: #11288
This commit appends hotplug devices (e.g., persistent volume)
to deviceInfos when `vfio_mod` is `vfio` and `cold_plug_vfio`
is set to one except `no-port`. For details, please visit the issue.
Signed-off-by: Hyounggyu Choi <Hyounggyu.Choi@ibm.com>
Adding:
"-object rng-random,id=rng0,filename=/dev/urandom -device
virtio-rng-pci,rng=rng0"
for confidential guests is not necessary as the RNG source cannot
be trusted and the guest kernel has the driver already disable as well.
Signed-off-by: Mikko Ylinen <mikko.ylinen@intel.com>
Switch imports to resolve:
```
SA1019: "github.com/opencontainers/runc/libcontainer/userns" is deprecated:
use github.com/moby/sys/userns
```
Signed-off-by: stevenhorsman <steven@uk.ibm.com>
In the latest oci-spec, the prestart hook is deprecated.
However, the docker & nerdctl tests failed when I switched
to one of the newer hooks which don't run at quite the same time,
so ignore the deprecation warnings for now to unblock the security fix
Signed-off-by: stevenhorsman <steven@uk.ibm.com>
We've been using the
github.com/containers/podman/v4/pkg/annotations module
to get cri-o annotations, which has some major CVEs in, but
in v5 most of the annotations were moved into crio (from 1.30)
(see https://github.com/cri-o/cri-o/pull/7867). Let's switch
to use the cri-o annotations module instead and remediate
CVE-2024-3056.
Signed-off-by: stevenhorsman <steven@uk.ibm.com>
- Detection of EROFS options in container rootfs
- Creation of necessary EROFS devices
- Sharing of rootfs with EROFS via overlayfs
Fixes: #11163
Signed-off-by: ChengyuZhu6 <hudson@cyzhu.com>
Some cni plugins will set the MTU of some routes, such as cilium will
modify the MTU of the default route. If the mtu of the route is not set
correctly, it may cause excessive fragmentation or even packet loss of
network packets. Therefore, this PR adds the setting of the MTU of the
route. First, when obtaining the route, if the MTU is set, the MTU will
also be obtained and set to the route in the guest.
Signed-off-by: Fupan Li <fupan.lfp@antgroup.com>
